Exploring Categorical Data
Summarize and display categorical data using frequency tables, bar charts, and pie charts, and draw conclusions from distributions.
Learning Objectives
- 1Construct and interpret frequency tables and relative frequency tables for categorical data
- 2Create and analyze bar charts and pie charts to represent categorical distributions
- 3Compare distributions of categorical variables across groups using segmented bar charts
- 4Identify misleading graphical representations and critique data displays
What Is Categorical Data?
Not all data is numerical. When researchers record eye color, political affiliation, blood type, or whether a patient responded to treatment, they are collecting categorical data — data that places individuals into groups or categories rather than measuring a quantity.
Categorical data answers the question Which group does this belong to? A response like "Yes / No / Maybe" or "Republican / Democrat / Independent" is categorical. So is a ZIP code — even though it looks like a number, arithmetic on ZIP codes is meaningless.
The statistical goal with categorical data is to summarize how many (or what proportion) of observations fall into each category.
Frequency Tables and Relative Frequencies
A frequency table lists each category alongside the count of observations in that category. A relative frequency table expresses those counts as proportions or percentages of the total.
| Blood Type | Frequency | Relative Frequency |
|---|---|---|
| O | 44 | 44% |
| A | 42 | 42% |
| B | 10 | 10% |
| AB | 4 | 4% |
| Total | 100 | 100% |
Relative frequencies allow comparisons across groups of different sizes — a critical skill when analyzing survey data with unequal sample sizes.
Think About
A study surveys 200 students at one school and 80 students at another. Why would you compare relative frequencies rather than raw counts when looking at favorite subject by school?
Bar Charts
A bar chart displays each category as a separate bar whose height (or length) represents frequency or relative frequency. Key features:
- Bars do not touch (categories are distinct, not continuous)
- The horizontal axis shows categories; vertical axis shows frequency or proportion
- Categories can appear in any order, though logical ordering (alphabetical, by size) aids interpretation
Segmented bar charts (also called stacked bar charts) show how a categorical distribution changes across groups. Each bar represents a group, and each segment shows the relative frequency of a category within that group.
When comparing two groups, AP exam questions often ask you to use relative frequencies (proportions) on the vertical axis — not raw counts — so that groups of different sizes can be fairly compared.
Think About
A segmented bar chart compares the preferred exercise type (running, swimming, cycling) of men and women. Both bars add to 100%. What does it mean if the bar segments are identical heights for both groups?
Pie Charts
A pie chart shows each category as a slice of a circle proportional to its relative frequency. Pie charts work best when:
- There are few categories (ideally fewer than six)
- The goal is showing part-to-whole relationships
Pie charts are generally harder to read precisely than bar charts because human perception is better at judging lengths than angles. On AP exams, bar charts are usually preferred for analysis.
Identifying Misleading Displays
Graphical displays can mislead. Watch for:
- Truncated y-axis: A bar chart whose vertical axis starts at a value other than zero exaggerates differences between bars
- 3D effects: Perspective distorts proportions in pie charts and bar charts
- Inconsistent scale: Bars of different widths or unevenly spaced axes distort visual comparison
- Missing labels: A graph without axis labels or a title is uninterpretable
❓Concept Check
A bar chart displays the market share of four phone brands. The y-axis begins at 15% instead of 0%. Brand A shows 18% and Brand B shows 20%. In the graph, Brand B's bar appears three times as tall as Brand A's. Is this display misleading? Why?
▸
Concept Check
A bar chart displays the market share of four phone brands. The y-axis begins at 15% instead of 0%. Brand A shows 18% and Brand B shows 20%. In the graph, Brand B's bar appears three times as tall as Brand A's. Is this display misleading? Why?
Yes, it is misleading. Because the y-axis is truncated (starts at 15%), the visual difference between bars exaggerates the actual difference. The true difference is only 2 percentage points (a 11% relative difference), but the truncated axis makes Brand B look dramatically larger. Always check that bar charts start at zero.
Comparing Distributions: Two-Way Tables
When categorical data involves two variables, we organize it in a two-way (contingency) table. The rows represent one variable, the columns represent another, and each cell contains the count for that combination.
Example: Grade level (9th, 10th, 11th, 12th) vs. preferred study time (morning, afternoon, evening).
From a two-way table, we can extract:
- Marginal distributions: The distribution of one variable ignoring the other (row or column totals)
- Conditional distributions: The distribution of one variable for a specific value of another
Comparing conditional distributions is the key to determining whether two categorical variables are associated. If the conditional distribution of preferred study time looks the same for every grade level, the variables appear independent. If they differ, there is an association.
Think About
In a two-way table of gender and career interest, 60% of women prefer STEM careers while 45% of men do. Is this evidence of an association between gender and career interest? What more would you need to conclude causation?
AP Exam Skill: Interpreting in Context
The AP Statistics exam consistently rewards responses that interpret statistical results in the context of the problem. For categorical data, this means:
- Name the variable(s) and categories explicitly
- State proportions with units ("45% of students surveyed preferred...")
- Make comparative statements when analyzing across groups
- Avoid causal language unless the data comes from a randomized experiment
Unit Summary
Categorical data organizes individuals into named groups. Frequency tables and relative frequency tables summarize categorical distributions numerically. Bar charts and pie charts display them visually. Two-way tables reveal relationships between two categorical variables, and comparing conditional distributions is the foundation of association analysis. A critical AP skill is recognizing and critiquing misleading graphical displays. Throughout all of this, interpretation in context — stating what the numbers mean for the specific situation — separates strong statistical reasoning from mere computation.
